Digitalization in Material Technology

M.Sc. Matthias Kahl, M.Sc. Sebastian Schramm

April 2019 - January 2022

Land Hessen

The project is part of a joint project in cooperation with the Institute of Materials Engineering/Metallic Materials, the Intelligent Embedded Systems Group, and the Control and System Theory Group of the University of Kassel. The Department of Measurement and Control has developed, examined, and showcased approaches for behavioral modelling in the area of additive manufacturing (SLM) in association with the aforementioned departments.

With the aim of an optimization and analysis of the selective laser melting (SLM) process the use of data-driven modeling approaches was examined in a first sub-project. The resulting models represent the causal relationship between various process control variables and the resulting part properties (hardness, density, porosity of the part). In order to explore a wide value range of the control variables, space-filling experiment-design methods were used and the resulting specimens characterized regarding their properties. Based on the acquired data, parametric and non-parametric models were generated using statistical and numerical methods and compared with regard to their prediction accuracy. Within the experimental limits, the obtained predictors allow the analysis of the influences of the process control variables on the part quality and the finding of suitable process windows leading to a desired part quality.

In a further sub-project, the data-driven spatio-temporal modeling of the laser metal deposition (LMD) process based on thermographic data was investigated. For this, methods were used facilitating the estimation of the parameters of partial difference equation models directly from the input and output data of the process. By doing so, a gray box model for predicting the temperature field of thin structures manufactured using the LMD process is identified from the thermographic measurement data.